Pages

Saturday, 9 May 2020

The art of the impossible



Suppose Keir Starmer were to turn into a horse. Even though such a transition is probably not contrary to Labour Party rules most of us would say it is impossible. Yet we could probably just about imagine Mr Starmer turning into a horse and a video whizz could certainly create a video of him appearing to do just that.

Pushing it just a little further, some might say that a Starmer to horse transition is at least conceivable by adding a few words about genes and genetic similarities within mammals. I can write about the idea and you can read about it even though it is impossible and lies well beyond the boundary of natural law.

Moving on. As many will know, there have been some destructive criticisms of the pandemic computer model supposedly used to guide government lockdown policy. Amateurish software would be a mild summary of this criticism.

However it does raise an interesting point because it may well be that the more extreme coronavirus model predictions were biologically impossible. That is to say, given the nature of the virus and a number of other factors, a much more severe pandemic may always have been a biological impossibility. It may have been as impossible as Mr Starmer changing into a horse and may also lie beyond the boundary of natural law.

Yet we tend not to look at things in this way. If future scenarios can be imagined or predicted by accredited people we tend to treat them as possibilities within known scientific laws. Yet they may not be. If we knew enough some future scenarios may not make sense. Within the laws of the natural world it may be that they cannot happen.

It is obvious enough that the pandemic model cannot tell us if certain scenarios are impossible. This in turn raises the possibility that pandemics are not computable. They are not a matrix of numbers, nor a matrix of equations, nor a matrix of quantifiable parameters. They are something else.

Even though we may measure the effect of pandemics and note that they follow certain general patterns, it may be that we cannot define fully what a pandemic actually is in the sense that we cannot tell what is possible and what is impossible. Part of that problem seems to be that those responsible for the models are able to imagine impossible scenarios and pass them on to policy makers. We are all able to imagine them. That does not make them possible.

8 comments:

Sam Vega said...

Yes, I think this is the age old problem of mistaking the map for the territory, with the added problem that a computer-generated map can depart from reality without us noticing it. Someone read the map, and now my children are bickering because they can't go out.

As for the Starmer example, you are quite right. A donkey may look like a horse in poor light, but it cannot turn into one.

Graeme said...

Whenever the Guardian applauds Starmer's alleged forensic skills at parliamentary question time, I always find myself asking how well his cabinet would have handled this pandemic. Forensic skills and gender equality programmes are no match for reality and viruses

Scrobs. said...

Surkeer is a politician via being a lawyer.

He won't understand Excel, and will never be able to grasp the nettle.

Only Scrobs can do this and I'm awaiting his call...

Doonhamer said...

That was the problem when hand held calculators came in. (I am old Father William)
People accepted the computed answer.
Common sense would demand that ball park figure had to be calculated by other means - mental arithmetic on rounded numbers, observing the real world, whatever.

A K Haart said...

Sam - you are right, yet mistaking the map for the territory ought to be a basic warning for anyone engaged in any kind of interpretation. So basic yet still missed.

Graeme - yes, forensic skills are no great advantage when it comes to inspiring people or even keeping their attention.

Scrobs - Surkeer is too posh for spreadsheets. He probably pronounces his name 'Suck Here'.

Doonhamer - I remember those first calculators too. Very useful in many respects but they also began to obscure what was really being done. They could be used for 'what if?' calculations. What if we assume the graph is still a straight line...

Nessimmersion said...

I wasn't aware that there was a horse in red dwarf.
Kryten on the other hand.

Ed P said...

Computer models have been misleading politicians for decades - they believe what they've already imagined, or worse, what they can use to push a hidden agenda.
Thus Mann's famous Hockey Stick, UEA's discredited climate model, etc, etc.
Now we have, "Instant Socialism by Influenza".
If only the great Corbyn was still the Glorious Leader! He waited for decades to inflict poverty on us all, then is cruelly replaced at the last minute by the robotic Kryten. Such a shame

A K Haart said...

Nessimmersion - surely Kryten is too funny. Maybe a blend of Kryten and Rimmer?

Ed - it amazes me that academic computer models haven't acquired a much more widespread negative reputation. The only conclusion that seems to makes sense is that they aren't really intended to work but merely intended to bring in funding.